Letters and numbers

⤓ Download the complete guidelines (PDF)

---

Transcription of Alphabetic and Numerical Signs

Since alphabet compositions and letterform variations are not consistent over time, some letters having emerged (such as w) while others tended to disappear (such as ſ and ƞ), we aimed to find a balanced approach between faithful rendering of the source and feasibility and reproducibility of the method across a broad scientific community.

General Principles

The transcription of alphabetic and numerical signs MUST follow the modern Latin alphabet. This ensures consistency and readability, as historical variations in letter shapes are too numerous and complex to classify systematically.

• Alphabetic signs MUST be transcribed using the modern Latin alphabet.
• Capital letters MUST be preserved.
• Small caps MUST be rendered as uppercase letters. Small caps are stylistic variations, not distinct characters, so they are normalized to uppercase for simplicity.
• Stylistic elements (e.g., italics, bold) MUST NOT be transcribed. These do not involve distinct characters.

Core Rule: The transcription MUST follow the letters as they appear in the source, with NO editorial corrections. This preserves the source text, which is critical for historical and linguistic analysis.

Ligatures

As ligatures represent graphical variations combining two letters into a single form, ligatures MUST be transcribed as separate letters.

In medieval documents, the ampersand (& [U+0026]) MUST be transcribed as is. This is based on the fact that this sign can bear additional abbreviation marks, such as macrons to indicate "etiam" in Latin.

In modern and contemporary documents, both "&" [U+0026] and "ß" [U+00DF] MUST be transcribed as is.

Ramist letters ("u"/"v", "i"/"j")

In medieval documents

In medieval documents, since the use of these signs does not distinguish between vocalic and consonantal uses, ramist letters are purely allographic variations. Therefore, Ramist letters (u/v, i/j) MUST be transcribed as i and u, respectively.

In modern and contemporary documents

In modern and contemporary documents, the transcription MUST follow modern usage (e.g., v for consonantal, u for vocalic).

In printed sources, the transcription MUST follow the printer’s usage. Printers’ conventions may differ, and respecting them ensures fidelity to the source.

Capital letters

Lowercase letters MUST NOT be normalized to uppercase.

If the distinction between uppercase and lowercase is unclear, the transcription MUST remain consistent.

Series of capital letters or emphasized letters, such as the one found in running title, MAY be transcribed as series of capital letter, depending on their inner variation.

Drop capitals and decorated letters MUST be transcribed as uppercase. and MUST be considered as not a part of the same text line as the line they semantically belong to: they should be identified as drop capitals during the image segmentation phase. Drop capitals often misalign with the baseline, so treating them separately avoids segmentation errors.

Oversized capitals expanding above the baseline, particularly at the beginning of paragraphs, MUST be transcribed, even if part of the letter is missing in the mask's polygon.

Digits and numerals

General Rules for Numerals

Numbers MUST be transcribed as they appear in the document, whether as Roman or Arabic numerals. This preserves the original format, which may be significant for historical or stylistic analysis.

• For Roman numerals, ramist standardization (u, v, i, j) MAY be applied following the rules specified above.
• Numbers written with superscript letters MUST follow the rules specified in the section dedicated to superscript abbreviations.
• The groups of numbers MUST be clearly separated, and the punctuation around the number MUST be retained where it exists.

In modern and contemporary documents, when the distinction between a capital I and a 1 is difficult to make (as can be the case in typewritten documents or in some manuscripts), we MAY follow the transcription that makes the most sense in the source.

Old-style roman numerals M and D (sometimes transcribed as "CIↃ" and "IↃ") MUST be transcribed as, respectively, M and D.

In modern documents, decimals MUST follow the convention used in the document (, or .). In case of a doubt, we recommend sticking to the system used more commonly in the rest of the document. If the source uses a symbol or a space to separate thousands, we should follow the report it in the transcription.